Research Advanced Computing Services Menu

• Compute
  • High Capacity 
  • Large Memory 
  • GPU-Enabled 
• Open OnDemand 
• Jupyter Notebooks 
• Scientific Software

• HPC Cluster Storage 
• Research Data Storage 
• MS OneDrive
• MS SharePoint

• Globus 
• Research Network
• Dryad Data Repository

• Facilitation Support
• Grant Development Support 
• Hardware Purchasing 
• Software Installation and Optimization 
• Data Support
• Data Security
• Access to National Supercomputing Resources 
• Public Cloud for Researchers 

High Performance Computing  

Compute

UO’s HPC cluster, Talapas, is comprised of the following combination of centrally funded open use resources, and researcher funded (condo) resources that are shared with all cluster users. Talapas has over 14,520 CPU cores, 129 TB of memory, and 89 GPUs. 

The following table summarizes the HPC resources offered by RACS:  

Open OnDemand (Web-Based Access) 

The RACS team supports a small cluster of Open OnDemand servers that provide a web-based graphical user interface to Talapas. Open OnDemand simplifies access to Talapas by offering a user-friendly, browser-based platform for both novice and experienced users. It allows users to view, edit, download, and upload files, manage and create job templates, and monitor resource usage on the cluster. Additionally, Open OnDemand supports interactive applications like remote desktops to cluster nodes, enabling users to access the power of the cluster in a familiar desktop environment. GUI-based software such as Matlab, JupyterLab/Notebook, and other scientific applications can also be launched directly, offering a seamless experience for running complex simulations, data analyses, or visualization workflows. This web-based approach eliminates the need for complex command-line interactions, making Talapas more accessible to users from various backgrounds, and enhancing productivity by streamlining the job submission and management process.

Jupyter Notebooks

Jupyter notebooks provide an interactive environment where you can write and execute code in small chunks, making it easy to test, iterate quickly, and maintain reproducibility. They seamlessly combine code, text, and visualizations, allowing you to document your work, create reports, and visualize data all in one place. This makes them ideal for prototyping, data analysis, and machine learning, with support for popular libraries like Pandas and Matplotlib. Jupyter is also language-agnostic, supporting multiple languages beyond Python, making it flexible for various projects. Plus, it enhances collaboration by allowing easy sharing and real-time teamwork. The UO’s Jupyter environment is available on Talapas via Open OnDemand.   

Scientific Software  

Numerous programming languages, compilers, and mathematical and scientific libraries are available cluster wide, including over 225 discipline-specific application packages and a large array of Python packages including machine learning specific packages such as TensorFlow, Keras, and PyTorch, among others. Additionally, RACS’ Research Software Engineer is available to help optimize applications. Researchers may request new software package installations or install software themselves in their home or project directories. 

Data Storage

HPC Cluster Storage 

Talapas provides over 3 Petabytes of IBM’s Elastic Storage System running Spectrum Scale parallel file system with access to local SSD scratch disk space for I/O intensive workloads; researchers are allocated a 250 GB individual home directory, and each research group is allocated 2 TB of project storage. Additionally, access to external storage via Microsoft OneDrive or SharePoint is available for backup/archive (see details below). The RACS team collaborates with the Information Security Office and researchers to ensure that appropriate security and compliance controls are in place for various research needs.   

Research Data Storage (coming soon!) 

Coming soon!  

Microsoft OneDrive 

The University of Oregon provides students, faculty and staff with 1 TB of file storage via Microsoft OneDrive. Microsoft OneDrive is approved for all data, however you must contact the Information Security Office BEFORE storing High Risk data in this location for the first time.   

Microsoft SharePoint 

The University of Oregon provides students, faculty and staff with 5 TB of file storage via Microsoft SharePoint. Microsoft SharePoint is approved for all data, however you must contact the Information Security Office BEFORE storing High Risk data in this location for the first time. 

Data Management

Globus 

Transferring data to and from Talapas is a common requirement in research computing and can be especially time-consuming when working with large datasets. The UO subscribes to Globus, which enables researchers fast, secure, and reliable data transfers between a desktop/laptop and Talapas, and access to data stored on a variety of national research clusters. Talapas provides dedicated Globus data transfer nodes connected via 100Gb/sec Ethernet to facilitate large data transfers. The Globus service at UO is hosted on our Science DMZ.  

Research Network 

The UO runs high-performance networks across campus and connects with our statewide network, Link Oregon, and the world. The campus core utilizes 100-Gigabit Ethernet with multiple 100-Gigabit links to Link Oregon. The UO is a member of Internet2, which operates a national network (400 GB/s) in support of U.S. research and education. The University is one of five founding members of Link Oregon, which operates a new facilities-based, resilient optical network (200+ Gb/s) across the state. Link Oregon connects to Internet2 (100 Gbps) and to our peer network in California, CENIC (100 Gbps). To support the University as well public and non-profit members in the lower Willamette Valley, Link Oregon maintains two redundant nodes in Eugene. Link Oregon also connects to high-speed Internet exchange points in Eugene (WIX), Portland (NWAX), and Seattle (SIX) for optimized regional performance.  

Science DMZ 

The University of Oregon’s Science DMZ is a specialized network infrastructure designed to facilitate the high-performance data transfer needs of the university's research community. Unlike traditional networks that prioritize general-purpose IT services, the UO Science DMZ is purpose-built to support the demanding requirements of data-intensive scientific applications, ensuring that researchers can move large datasets efficiently and securely.  

Security and Monitoring  

Security within the UO Science DMZ is tailored to maintain high performance while ensuring the integrity of research data. Custom security policies are enforced to protect sensitive information without compromising data transfer speeds. Performance testing is carried out using perfSONAR nodes, ensuring the network remains efficient under heavy loads.  

Integration and Support  

Based on the Science DMZ model developed by the U.S. Department of Energy's Energy Sciences Network (ESnet), the UO Science DMZ integrates seamlessly with the broader university network while maintaining distinct pathways for research data. This design ensures that academic and administrative traffic does not interfere with research operations, while still allowing for efficient resource sharing across campus.  The UO Science DMZ is continuously enhanced to keep pace with the evolving needs of its research community. 

Dryad Data Repository  

The University of Oregon is an institutional member of Dryad, a non-profit, community-governed data repository. Any UO researcher can publicly share their research data in Dryad at no cost. Sharing data in Dryad complies with funder and publisher data-sharing policies. For more information, see the UO Libraries research guide on Dryad Data Repository and/or contact: DataServices@uoregon.edu.    

Consulting, Outreach and Education   

Research Facilitation (coming soon!) 

RACS is hiring a Research Computing Facilitator to assist researchers throughout the research process. They will provide a wide range of support including identifying resources and funding opportunities to identifying collaborators, education for staff and students, and directly participating on research teams. Facilitators work with investigators, coordinators, staff, and students to facilitate research and reduce barriers to the research process.  

Grant Development Support  

RACS offers assistance to researchers who are seeking funding in the form of external grants from research centers and federal agencies, such as the National Science Foundation (NSF) and the National Institutes of Health (NIH). RACS can provide letters of support for grant proposals from researchers who have utilized RACS resources in their research projects. To discuss letters of support, data management, facilities and resources documents, or hardware quotes for your grant, please contact us. 

If you are submitting a grant proposal for research that will utilize RACS resources, you will likely require some background information about RACS’ resources and facilities.

Acknowledging RACS in your publications  

All forms of publication, including web pages, resulting from work done using RACS resources should include the following citation:  

The authors acknowledge Research Advanced Computing Services (RACS) at the University of Oregon for providing computing resources that have contributed to the research results reported within this publication. URL: https://racs.uoregon.edu.

Condo and Hardware Purchasing 

The condo model allows researchers to purchase dedicated computing resources tailored to their specific needs. The RACS team provides support with testing, configuring, and installing these systems to interoperate seamlessly with the Talapas cluster. Additionally, RACS assists with hardware selection for grant proposals and can provide supporting quotes and necessary documentation during the proposal process. RACS also handles all administrative tasks related to condo node management as part of its standard services, ensuring full support throughout the process of purchasing and maintaining condo resources.

Software Installation and Optimization 

Numerous programming languages, compilers, and mathematical and scientific libraries are available cluster wide, including over 225 discipline-specific application packages and a large array of Python packages including machine learning specific packages such as TensorFlow, Keras, and PyTorch, among others. Additionally, RACS’ Research Software Engineer is available to help optimize applications. Researchers may request new software package installations or install software themselves in their home or project directories. 

Data Support

Data is a critical component of research, and we offer comprehensive support for a several data storage platforms to meet the diverse needs of our users on the cluster. Whether you're working with our high-performance parallel file system, cloud-based storage solutions, or traditional network-attached storage, our infrastructure is designed to handle large datasets efficiently. We provide seamless integration with popular storage services, ensuring that users can easily access, store, and manage their research data, regardless of the platform. Our team also offers guidance and technical support to help optimize data workflows and ensure data security and integrity across the entire research lifecycle.

Data Security

For more information about data classification and security in research, see the Information Security Office's website. 

Access to National Supercomputing Resources 

In addition to Talapas, some researchers may need to access resources at national supercomputing centers. ACCESS is an NSF-funded platform that provides access to a variety of computational and storage resources for US-based university and non-profit researchers at no cost. Members of the RACS team are part of the Campus Champions program, designed to help researchers determine which national resources are most appropriate for their computational needs and help prepare time grant proposals needed to obtain access.  

Public Cloud for Researchers 

The University of Oregon has service agreements with two public cloud services providers: Amazon Web Services (AWS) and Microsoft Azure. Cloud providers are best suited for on-demand virtual machines, website hosting, networking, cloud storage, mobile development, and other specialized use-cases. Information Services (IS) / Research Advanced Computing Services (RACS) provides consulting support for researchers interested in using public cloud services. Our teams can help you determine if cloud is a good fit for your computational requirements, plan for and manage costs of cloud computing, and gain access to cloud computing resources for research and research-related instruction.